Magic will certainly have a pervasive impact. I think without defining how prevalent magic is and how it works it will be impossible to answer exactly how impactful it might be. I think much of the extrapolation often depends on several assumptions I find unreasonable, first that high level magic can be made prevalent and second that any magic and especially high level magic has no side effects simply because none are listed in things like spell descriptions. And I think that magic especially powerful magic is likely to have side effects and that just because we don't really know how it works or just because those potential side effects aren't described doesn't mean that they aren't there.
So first, I think the most impactful magic will be the ones that are available at low level because those are the ones that will be most pervasive and sustainable. I don't believe that high level spell casters are something that most societies can produce in significant quantities and though even a few of them will have a big impact, it won't be the most impactful simply because of its rarity and its unsustainability. The rarer the high level caster the more likely their works will break or be destroyed before another high level caster comes along with the willingness to repair them.
And to that we should now append that again, it really matters how magic works. In 3e for example almost all permanent magic involves an expenditure of experience points, which is far more valuable than the expenditure of gold or even time, as it not only involves the spellcaster spending some of their past life, but sells off their future as well by setting back their development. So things that cost XP are going to be limited according to the availability of that resource, which for most NPCs is going to be a very dear resource indeed. In 1e almost all permanent magic involves risking the permanent loss of health in the form of Constitution loss. Again, this is a very dear resource indeed. And this is just the part of the magic we have some knowledge of how it works. We know vaguely that that 'gold' in 3e actually abstracts out the expenditure of rare magical consumables, such as gems or monster parts. But without knowing for example how many Griffins must be slain to create a magic item that requires Griffin parts, we don't really know anything about the long term impact or the scalability of the 'technology'. It could be that high demand would drive magical beasts to extinction, eventually ending the technology.
Consider the case of the zombie powered grist mill. We know that zombies can power a grist mill. We have no idea how quickly a zombie body will fall apart if eternally condemned to turn a grist mill. We do know that zombies seem to have no ability to heal on their own, and that I would think means that you'd need some sort of zombie servicing technician capable of healing zombies from the damage of decay and wear and tear that goes below the level of hit point injuries but which would certainly accumulate. And so now you must not only postulate someone is creating zombies to run grist mills, but also that someone with the magic to repair zombies is willing to maintain them and that communities are OK with grandma being zombified and turned into a component of a grist mill and that you can do all of that with no magical or spiritual side effects. Maybe, and as a DM you could say that indeed that does happen in your game, but I'm skeptical that if realism is the goal here that it will be simple or necessarily vastly superior to having a few donkeys turn the mill.
Or in short not only must most impactful magic be prevalent and accessible, it must also be permanent, cheap, and renewable.
Finally, magic if it has a negative impact will almost certainly have a strong push back. If people don't feel magic on the whole is making life better, they will justly fear and hate those that abuse it or even practice it. A good example is mind controlling magic. A spell like 'Charm Person' as some have pointed out would have a huge impact on a setting. But the impact is so large and so wholly negative that it's likely to just result in everyone known to be or believed to be capable of casting 'Charm Person' murdered in their sleep, and society feeling that it had not done ill by ridding the world of such a monstrous power. That is a huge impact on the world, but not necessarily the sort that I think the OP was thinking of.
So additionally, the impact of magic must either be inescapable or else it must be one that society will tolerate and allow or else the actual impact can only be great in the sort of persistent resistance to and defenses against magic it creates. And further, it's impact must be measured in terms of some sort of balance or demographics which, this being a fantasy game, I don't feel amiss in describing as not less than the conflict between good and evil. Enriching plants via plant growth indeed could have a potentially large impact on society, comparable to scientific farming methods in impact (though on its own not as great, presumably much of the rest could also be magically replicated). But that impact is only actually realized if the number of spell casters enriching plant growth exceeds those stunting plant growth by the equally accessible application of negative magic, and that in turn will depend on unanswered questions we have about the nature of spell casting - who can learn it, how widespread it can be, how tied to some sort of cosmic balance it is, and so forth.
I haven't rigorously looked at an SRD for the purpose of addressing this question (and I'll probably do that now that the thread is suggesting I do so) but I'd second a vote for something as simple as 'Mending' as addressing all the issues that I brought up and therefore likely making a pretty big impact on society. It at the very least suggests that being a magical tinkerer and fixing things is a viable trade that people are likely to accept as long as you confine yourself to something that they can understand and appreciate like that, and it's economic impact while subtle is I think going to be rather powerful if it becomes pervasive if you think about how many "broken windows" it saves and how many man days of labor to replace them it saves at a pretty trivial cost.
Endure Elements has certainly had an impact in our current campaign as the party travels through the steaming jungle, but its impact on the larger world again depends on demographics. Before you can really use it to replace skills like Endurance and Survival, you need to be able to cast it on a significant portion of the work force - and it's not really clear that enough spellcasters exist to do that. Obviously, you could postulate a world where routinely common laborers were low level spellcasters in order to take advantage of Endure Elements, and assuming they could also and at the same time acquire the other skills they needed for labor, then that might work. But it's not at all clear that D&D or most RPGs really do endorse the idea of everyone being a 1st level wizard, and you have to wonder if there wouldn't be subtle consequences to that like everyone has a bad Fortitude save that would make the society fragile in the long run. And that depends on the demographics of commoner in your world.